diff --git a/torchrl/data/datasets/atari_dqn.py b/torchrl/data/datasets/atari_dqn.py
index 6193bd96afa..42dc35bfac5 100644
--- a/torchrl/data/datasets/atari_dqn.py
+++ b/torchrl/data/datasets/atari_dqn.py
@@ -60,7 +60,7 @@ class AtariDQNExperienceReplay(BaseDatasetExperienceReplay):
         root (Path or str, optional): The AtariDQN dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/atari`.
+            `~/.cache/torchrl/atari`.atari`.
         num_procs (int, optional): number of processes to launch for preprocessing.
             Has no effect whenever the data is already downloaded. Defaults to 0
             (no multiprocessing used).
diff --git a/torchrl/data/datasets/d4rl.py b/torchrl/data/datasets/d4rl.py
index a1567aa0385..0189678dd1d 100644
--- a/torchrl/data/datasets/d4rl.py
+++ b/torchrl/data/datasets/d4rl.py
@@ -106,7 +106,7 @@ class D4RLExperienceReplay(BaseDatasetExperienceReplay):
         root (Path or str, optional): The D4RL dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/d4rl`.
+            `~/.cache/torchrl/atari`.d4rl`.
         download (bool, optional): Whether the dataset should be downloaded if
             not found. Defaults to ``True``.
         **env_kwargs (key-value pairs): additional kwargs for
diff --git a/torchrl/data/datasets/gen_dgrl.py b/torchrl/data/datasets/gen_dgrl.py
index ddde2e9d65d..672ec8465bf 100644
--- a/torchrl/data/datasets/gen_dgrl.py
+++ b/torchrl/data/datasets/gen_dgrl.py
@@ -60,7 +60,7 @@ class GenDGRLExperienceReplay(BaseDatasetExperienceReplay):
             dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/gen_dgrl`.
+            `~/.cache/torchrl/atari`.gen_dgrl`.
         download (bool or str, optional): Whether the dataset should be downloaded if
             not found. Defaults to ``True``. Download can also be passed as ``"force"``,
             in which case the downloaded data will be overwritten.
diff --git a/torchrl/data/datasets/minari_data.py b/torchrl/data/datasets/minari_data.py
index 113c2595b77..126d0dadc93 100644
--- a/torchrl/data/datasets/minari_data.py
+++ b/torchrl/data/datasets/minari_data.py
@@ -66,7 +66,7 @@ class MinariExperienceReplay(BaseDatasetExperienceReplay):
         root (Path or str, optional): The Minari dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/minari`.
+            `~/.cache/torchrl/atari`.minari`.
         download (bool or str, optional): Whether the dataset should be downloaded if
             not found. Defaults to ``True``. Download can also be passed as ``"force"``,
             in which case the downloaded data will be overwritten.
diff --git a/torchrl/data/datasets/openx.py b/torchrl/data/datasets/openx.py
index 01f5fdf98ce..c5b5dc9e8fc 100644
--- a/torchrl/data/datasets/openx.py
+++ b/torchrl/data/datasets/openx.py
@@ -123,7 +123,7 @@ class for more information on how to interact with non-tensor data
         root (Path or str, optional): The OpenX dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/openx`.
+            `~/.cache/torchrl/atari`.openx`.
         streaming (bool, optional): if ``True``, the data won't be downloaded but
             read from a stream instead.
 
diff --git a/torchrl/data/datasets/roboset.py b/torchrl/data/datasets/roboset.py
index f967ddfa1fc..1a83c302860 100644
--- a/torchrl/data/datasets/roboset.py
+++ b/torchrl/data/datasets/roboset.py
@@ -57,7 +57,7 @@ class RobosetExperienceReplay(BaseDatasetExperienceReplay):
         root (Path or str, optional): The Roboset dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/roboset`.
+            `~/.cache/torchrl/atari`.roboset`.
         download (bool or str, optional): Whether the dataset should be downloaded if
             not found. Defaults to ``True``. Download can also be passed as ``"force"``,
             in which case the downloaded data will be overwritten.
diff --git a/torchrl/data/datasets/vd4rl.py b/torchrl/data/datasets/vd4rl.py
index 7290a714155..f1b2cc673cc 100644
--- a/torchrl/data/datasets/vd4rl.py
+++ b/torchrl/data/datasets/vd4rl.py
@@ -63,7 +63,7 @@ class VD4RLExperienceReplay(BaseDatasetExperienceReplay):
         root (Path or str, optional): The V-D4RL dataset root directory.
             The actual dataset memory-mapped files will be saved under
             `<root>/<dataset_id>`. If none is provided, it defaults to
-            ``~/.cache/torchrl/vd4rl`.
+            `~/.cache/torchrl/atari`.vd4rl`.
         download (bool or str, optional): Whether the dataset should be downloaded if
             not found. Defaults to ``True``. Download can also be passed as ``"force"``,
             in which case the downloaded data will be overwritten.
diff --git a/torchrl/data/map/query.py b/torchrl/data/map/query.py
index ff0fb4dfe24..6c4c2f9e0e2 100644
--- a/torchrl/data/map/query.py
+++ b/torchrl/data/map/query.py
@@ -80,12 +80,12 @@ class QueryModule(TensorDictModuleBase):
             If a single ``hash_module`` is provided but no aggregator is passed, it will take
             the value of the hash_module. If no ``hash_module`` or a list of ``hash_modules`` is
             provided but no aggregator is passed, it will default to ``SipHash``.
-       clone (bool, optional): if ``True``, a shallow clone of the input TensorDict will be
+        clone (bool, optional): if ``True``, a shallow clone of the input TensorDict will be
             returned. This can be used to retrieve the integer index within the storage,
             corresponding to a given input tensordict. This can be overridden at runtime by
             providing the ``clone`` argument to the forward method.
             Defaults to ``False``.
-    d
+
         Examples:
             >>> query_module = QueryModule(
             ...     in_keys=["key1", "key2"],
@@ -106,6 +106,7 @@ class QueryModule(TensorDictModuleBase):
             >>> # The last three pairs of key1 and key2 have at least one mismatching value
             >>> assert res["index"][1] != res["index"][2]
             >>> assert res["index"][2] != res["index"][3]
+
     """
 
     def __init__(
diff --git a/torchrl/data/replay_buffers/replay_buffers.py b/torchrl/data/replay_buffers/replay_buffers.py
index 4e0ee36cd4a..e2f2918b3dc 100644
--- a/torchrl/data/replay_buffers/replay_buffers.py
+++ b/torchrl/data/replay_buffers/replay_buffers.py
@@ -99,6 +99,7 @@ class ReplayBuffer:
             is used with PyTree structures (see example below).
         batch_size (int, optional): the batch size to be used when sample() is
             called.
+
             .. note::
               The batch-size can be specified at construction time via the
               ``batch_size`` argument, or at sampling time. The former should
@@ -108,6 +109,7 @@ class ReplayBuffer:
               incompatible with prefetching (since this requires to know the
               batch-size in advance) as well as with samplers that have a
               ``drop_last`` argument.
+
         dim_extend (int, optional): indicates the dim to consider for
             extension when calling :meth:`extend`. Defaults to ``storage.ndim-1``.
             When using ``dim_extend > 0``, we recommend using the ``ndim``
@@ -128,6 +130,7 @@ class ReplayBuffer:
                     >>> for d in data.unbind(1):
                     ...     rb.add(d)
                     >>> rb.extend(data)
+
         generator (torch.Generator, optional): a generator to use for sampling.
             Using a dedicated generator for the replay buffer can allow a fine-grained control
             over seeding, for instance keeping the global seed different but the RB seed identical
@@ -582,6 +585,7 @@ def register_save_hook(self, hook: Callable[[Any], Any]):
 
         .. note:: Hooks are currently not serialized when saving a replay buffer: they must
             be manually re-initialized every time the buffer is created.
+
         """
         self._storage.register_save_hook(hook)
 
@@ -926,8 +930,8 @@ class PrioritizedReplayBuffer(ReplayBuffer):
             construct a tensordict from the non-tensordict content.
         batch_size (int, optional): the batch size to be used when sample() is
             called.
-            .. note::
-              The batch-size can be specified at construction time via the
+
+            .. note:: The batch-size can be specified at construction time via the
               ``batch_size`` argument, or at sampling time. The former should
               be preferred whenever the batch-size is consistent across the
               experiment. If the batch-size is likely to change, it can be
@@ -935,6 +939,7 @@ class PrioritizedReplayBuffer(ReplayBuffer):
               incompatible with prefetching (since this requires to know the
               batch-size in advance) as well as with samplers that have a
               ``drop_last`` argument.
+
         dim_extend (int, optional): indicates the dim to consider for
             extension when calling :meth:`extend`. Defaults to ``storage.ndim-1``.
             When using ``dim_extend > 0``, we recommend using the ``ndim``
@@ -1051,6 +1056,7 @@ class TensorDictReplayBuffer(ReplayBuffer):
             construct a tensordict from the non-tensordict content.
         batch_size (int, optional): the batch size to be used when sample() is
             called.
+
             .. note::
               The batch-size can be specified at construction time via the
               ``batch_size`` argument, or at sampling time. The former should
@@ -1060,6 +1066,7 @@ class TensorDictReplayBuffer(ReplayBuffer):
               incompatible with prefetching (since this requires to know the
               batch-size in advance) as well as with samplers that have a
               ``drop_last`` argument.
+
         priority_key (str, optional): the key at which priority is assumed to
             be stored within TensorDicts added to this ReplayBuffer.
             This is to be used when the sampler is of type
@@ -1085,6 +1092,7 @@ class TensorDictReplayBuffer(ReplayBuffer):
                     >>> for d in data.unbind(1):
                     ...     rb.add(d)
                     >>> rb.extend(data)
+
         generator (torch.Generator, optional): a generator to use for sampling.
             Using a dedicated generator for the replay buffer can allow a fine-grained control
             over seeding, for instance keeping the global seed different but the RB seed identical
@@ -1394,6 +1402,7 @@ class TensorDictPrioritizedReplayBuffer(TensorDictReplayBuffer):
             construct a tensordict from the non-tensordict content.
         batch_size (int, optional): the batch size to be used when sample() is
             called.
+
             .. note::
               The batch-size can be specified at construction time via the
               ``batch_size`` argument, or at sampling time. The former should
@@ -1403,6 +1412,7 @@ class TensorDictPrioritizedReplayBuffer(TensorDictReplayBuffer):
               incompatible with prefetching (since this requires to know the
               batch-size in advance) as well as with samplers that have a
               ``drop_last`` argument.
+
         priority_key (str, optional): the key at which priority is assumed to
             be stored within TensorDicts added to this ReplayBuffer.
             This is to be used when the sampler is of type
@@ -1431,6 +1441,7 @@ class TensorDictPrioritizedReplayBuffer(TensorDictReplayBuffer):
                     >>> for d in data.unbind(1):
                     ...     rb.add(d)
                     >>> rb.extend(data)
+
         generator (torch.Generator, optional): a generator to use for sampling.
             Using a dedicated generator for the replay buffer can allow a fine-grained control
             over seeding, for instance keeping the global seed different but the RB seed identical
@@ -1669,6 +1680,7 @@ class ReplayBufferEnsemble(ReplayBuffer):
             Defaults to ``None`` (global default generator).
 
             .. warning:: As of now, the generator has no effect on the transforms.
+
         shared (bool, optional): whether the buffer will be shared using multiprocessing or not.
             Defaults to ``False``.
 
diff --git a/torchrl/data/rlhf/utils.py b/torchrl/data/rlhf/utils.py
index 26fb710f94c..bbde6761f4a 100644
--- a/torchrl/data/rlhf/utils.py
+++ b/torchrl/data/rlhf/utils.py
@@ -198,6 +198,7 @@ class RolloutFromModel:
             batch_size=torch.Size([4, 50]),
             device=cpu,
             is_shared=False)
+
     """
 
     EOS_TOKEN_ID = 50256